smbd: lib/util_sock.c:(436) / read failure for 4

[Leon Allen]

Here is a short excerpt from out hpux 11i syslog file:

Mar 21 09:35:00 cis1 su: + tty?? root-root
Mar 21 09:43:42 cis1 smbd[14110]: [2005/03/21 09:43:42, 0] lib/util_sock.c:(436)
Mar 21 09:43:42 cis1 smbd[14110]: read_data: read failure for 4. Error = Connection reset by peer
Mar 21 09:43:53 cis1 smbd[14131]: [2005/03/21 09:43:53, 0] lib/util_sock.c:(436)
Mar 21 09:43:53 cis1 smbd[14131]: read_data: read failure for 4. Error = Connection refused
Mar 21 09:45:01 cis1 su: + tty?? root-root
Mar 21 09:50:00 cis1 su: + tty?? root-root
Mar 21 10:00:01 cis1 su: + tty?? root-root
Mar 21 10:15:03 cis1 su: + tty?? root-root
Mar 21 10:17:50 cis1 above message repeats 2 times
Mar 21 10:22:01 cis1 smbd[14160]: [2005/03/21 10:22:01, 0] lib/util_sock.c:(436)
Mar 21 10:22:01 cis1 smbd[14160]: read_data: read failure for 4. Error = Connection reset by peer
Mar 21 10:30:00 cis1 su: + tty?? root-root
Mar 21 10:35:00 cis1 su: + tty?? root-root
Mar 21 10:45:01 cis1 su: + tty?? root-root
Mar 21 10:50:00 cis1 su: + tty?? root-root
Mar 21 11:00:01 cis1 su: + tty?? root-root
Mar 21 11:15:33 cis1 smbd[15476]: [2005/03/21 11:15:33, 0] lib/util_sock.c:(436)
Mar 21 11:15:33 cis1 smbd[15476]: read_data: read failure for 4. Error = Connection reset by peer
Mar 21 11:15:00 cis1 su: + tty?? root-root
Mar 21 11:17:55 cis1 above message repeats 2 times
Mar 21 11:30:01 cis1 su: + tty?? root-root
Mar 21 11:31:09 cis1 smbd[15452]: [2005/03/21 11:31:09, 0] lib/util_sock.c:(436)
Mar 21 11:31:09 cis1 smbd[15452]: read_data: read failure for 4. Error = Connection timed out
Mar 21 11:35:00 cis1 su: + tty?? root-root
\It seems to coincide with citical application (Chris HR Payrol) terminating abriptly.

Any Clues? Any one seen this before? I've done a search, and haven't come up with anything useful yet.

My gut feel is a network error. How can I confirm this? Is there tracinr / logging I can turn on to thow some light on this?

More importantly, how do I fix it?

Many thanks.

[harry d brown jr]

I have never seen those errors, but take a look at this:

http://lists.samba.org/archive/samba/2002-September/051952.html

live free or die
harry d brown jr

[Leon Allen]

Thanks for the reply.

Something happens on our server (a network / tcp failure??). There are two symptoms:

1) samba logs an error - indicating the client has disconnected
2) Our HR application falls over (this is the killer, and the HR app does not use samba)

These are the only evidence / symptoms of the problem I can find.

How can I find out more to diagnose the problem? Where does the problem really lie: HP server, network, tcp/ip, client, or app?

I suspect a load (ie busy) induced failure in the network (tcp stack?) but am likely to be completely wrong.
Is there some logging I can turn on / examine.

ps I will check out the link you suggested now :-)

Many thanks!

Leon

[Leon Allen]

Re: the suggested link - yes, a result of the little research I had done I had set those parameters as per the article (:-(